View
8
Download
0
Category
Preview:
Citation preview
NASA/TM-2000-209603
AFDD/TR-00-A-009
Evaluation of Helmet Mounted Display Alerting Symbology
Joe De Maio
Aeroflightdynamics Directorate
U.S. Army Aviation and Missile Command
Ames Research Center, Moffett Field, California
National Aeronautics and
Space Administration
Ames Research Center
Moffett Field, California 94035-1000
September 2000
https://ntrs.nasa.gov/search.jsp?R=20010047013 2020-05-26T21:47:25+00:00Z
NASA Center for AeroSpace Information7121 Standard Drive
Hanover, MD 21076-1320
(301) 621-0390
Available from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
(703) 487-4650
EVALUATION OF HELMET MOUNTED DISPLAY ALERTING
SYMBOLOGY
Joe De Maio
Aeroflightdynamics Directorate
U.S. Army Aviation and Missile CommandAmes Research Center
INTRODUCTION
The present research is aimed at developing tools to support the NASA Safe All-
weather Flight Operations Research (SAFOR) program. The goal of our part of
this program is to make dramatic reductions in the rate and severity of civil
rotorcraft accidents through reduction in pilot error. This research is also
supported by a cooperative agreement between the Army AeroflightdynamicsDirectorate and NASA Ames Research Center.
A major factor in pilot error is the failure to apprehend critical information or to
integrate isolated facts into a coherent concept that can guide correct behavior.
An important tool for improving pilot situation awareness is the helmet mounted
display. In its earliest implementations, the helmet display was used primarily
as a weapon pointing device. With the addition of sensor imagery it came to be
an aid to flight in limited visibility.
The benefit from a helmet mounted display increases when the pilot has a
lessened requirement to look back into the cockpit. Ultimately it is desirable to
make the helmet display the primary flight instrument. In fact the Army's RAH-
66 Comanche will have the helmet display as its primary flight display. The
helmet display needs to be well integrated with the panel instruments when
used in this way. It needs to present as much information as possible without
being cluttered or obscuring the out-the-window view. It also needs to permit an
easy transition back into the cockpit when this is needed.
Pilots face purely mechanical problems in shifting from the helmet display to the
panel display. They must refocus and reconverge their eyes from infinity to
under 31 inches (Hawkins, 1997, p 246). They must also adapt to a different
display brightness. These problems compound the biggest problem, that of
switching attention from the out-the-window task to the in-cockpit task.
Switching from one task to another can be a slow process. It can take as much astwo to three seconds to switch from one attention demanding task to another
demanding task. Switching time can be reduced substantially by effective
alerting (De Maio, 1976).
The research used two approaches to increasing the effectiveness of alerts. One
was to increase the ability of the alert to attract attention by using the entire
display surface. The other was to include information about the required
response in the alert itself.
In the full screen alert, a flashing alert symbol appears at the bottom of the
helmet display. In addition to this flashing symbol, all of the nominal symbology
on the display flashes until the pilot were to acknowledge the alert. Stimulating
a large area of the retina with the flashing alert can stimulate the part of the
visual system that operates without attention. This ambient visual system
(Hennessey and Sharkey, 1997) is primarily involved in postural control and
orienting, but it may also serve to alert the viewer to movement in the periphery
of the visual field. This system is relatively insensitive to the attentional demand
of the primary task. Thus ambient alerting can pull attention from that task even
when alerts are rare and the primary task has a high attentional demand.
Adding task information to the alert should aid the pilot in transitioning from
the flying task to the alert response task. It should not affect the alerting quality
of the display. The information about the alert task allows the pilots to begin
thinking about the task even while their eyes are still directed at the helmet
display. Partial information about the response has been shown to speed
responding in choice reaction time tasks (Leuthold et al, 1996). In addition, the
information manipulation works by improving the visual momentum between
the helmet display and the panel display. Visual momentum refers to the ease
with which the viewer can transfer attention smoothly from one display to the
other (Woods et al, 1987).
The helmet display evaluation was conducted along with a test of workload and
situation awareness measures. The present report contains a summary of the test
method. The details are presented in a separate report (De Maio and Hart, 1999).
2
METHOD
Apparatus
Helicopter Simulation - The investigation was conducted using the NASA Ames
Research Center's six-degree-of-freedom vertical motion simulator (VMS) with a
rotorcraft cockpit. The VMS is unique among flight simulators in its large range
of motion. This large motion capability provides high quality flight cues to the
pilot.
A simulated rotorcraft cabin, the RCAB, was configured as a single-pilot cockpit
with a four-window computer generated display, consisting of three forward
view, CRT displays, spanning 27 ° X 147 ° and one CRT chin window on the right
side (26 ° X 22°). The out-the-window imagery was generated using an Evans
and Sutherland ESIG 3000 image generator.
The primary inputs to the motion base are the aircraft translational and
rotational accelerations calculated by the math model for the pilot position.
Appendix A contains a summary of the motion gains. The simulated aircraft was
a UH-60A, Black Hawk. Rotor, engine, and transmission sounds were simulated.
Conventional helicopter controls were used. The stick-to-visual throughput time
delay was approximately 72 msec. In addition there was a 20 msec math model
cycle time.
Panel instruments were displayed on two 14 in diagonal color CRTs. The right
CRT displayed generic, basic flight instruments (see Figure 1). The left CRT
displayed a moving map of the visual database (see Figure 2). The map showed
major terrain features and major roads, drawn in light blue. The planned course
was displayed in red. A compass rose was displayed in the upper right hand
comer. A gray square overlaid the high resolution area at the center of the data
base. In the visual database, this area was a high definition rendering of a small
village, which was not represented on the map. A digital range indicator in thelower left comer indicated the size of the displayed area in nautical miles. When
the alert task was presented, the required input and the pilot's response were
overlaid at the bottom of the map. The helmet mounted display system
consisted of the helmet, helmet display unit and head position sensing system of
the AH-64 integrated helmet and display sight system.
Helmet Display Symbology - Helmet display symbology (see Figure 3) was
based on the AH-64's pilot night vision system (PNVS), cruise mode symbology.
This symbology includes a compressed 120 ° compass at the top of the display,
digital torque and airspeed on the left, digital altitude and analog, radar altitude
and vertical speed on the right. A dashed line gives an indication of pitch and
roll, referenced to the display frame. A diamond indicating the position of the
aircraft's nose is the only head slaved PNVS symbology. Symbology was added
for the course director indicator (CDI), Waypoint, and alert displays. In the
simulator, altitude was above ground level for both digital and analog displays.
Alert Task Symbology - There were five alert type conditions. A No-alert
condition provided a flying performance baseline. Alert flash (Localized andFull-Screen) was crossed with alert information content (No-Info and Partial-
Info) to yield four experimental conditions (see Table 1). The alert was presented
3
solely on the helmet mounted display, with a digit entry task to simulate the
procedural response presented on the panel.
•-.'_ 320 I // .. t_¢ 40 "C. ---_ 4. " 9 ' G-
1_.240 KNOT5 12%_ G--7 LT 3_
¢;._.200 _.O.,\t" _, \+m
\xx_ll f_it, ,\_l'31_'_J/,.C' 90 JlO -. lu L'0 ".
80 r',o .,. " '_ CLIMB
•,j_, 30 T,_, rr ,_ .,, 30 -L:.._ o_ I | I -
% PERCENt ,v, _" _
C/ '10 .. ?0 ," DI " 10 20/ 3'J • -\ "_x"//ttt_O \" T t_ ta_ 1_,\',
%'_',1/1' I
j.,./ 1 I:
_ ''''dni4_' so c_ "/,,/e (. r_x_.- _..iO ALT _ ...,.
Figure 1. Simulated instrument panel. Instruments consisted of
white and colored graphics on a black background. Color scheme
has been revised for better printing.
In all Alert Task conditions, a flashing letter was presented in the bottom center
of the display, the position corresponding to the TADS field of view box (see
Figure 4). The letter was approximately 3 ° tall. In the Localized alert condition,
this symbol constituted the entire alert. In the Full-Screen alert condition, all
symbology flashed. Flashing alternated between the center and periphery of the
display, that is, when the center brightness was high, the periphery was low and
vice versa. Central symbology consisted of the horizon line, nose diamond, alert
letter, and the waypoint when it was pre.sent. All other symbology was
peripheral. "High" brightness was the brightness set by the pilot. "Low"
brightness voltage was 30% of high. This design ensured that some symbology
would always be fairly bright and that no symbology would ever be completely
off. Flash rate was set to be noticeable but not disturbing, at three Hz with a
linear ramp up and down. Three alert symbols were used. In the No-
Information condition, an upper case "N" indicated an alert. In the Part-Information condition, an '%" indicated an alert and that numbers were to be
entered left-to-right, while an "R" indicated right-to-left entry.
\\
\
Figure 2. Map display. Terrain contours consisted of colored lines
on a black background. Roads were teal. Planned route was red.
Color scheme has been revised for better printing.
Navigation Symbology - There were three navigation display conditions. A
baseline "Visual" navigation condition used the basic PNVS symbology without
the waypoint symbol on the compass display. A "Waypoint" condition used a
waypoint marker superimposed on the visual scene. The "CDI" condition
incorporated symbols into the compass display indicating bearing to waypoints
and course deviation. Data were collapsed across these conditions for the alert
display analysis.
Table 1. Alert conditions.
No Alert No PartialInformation Information
No Alert Baseline
Localized Alert X X
Full-Screen Alert X X
The two aided displays included an arrival time clock in the upper right that
showed the pilot's instantaneous arrival time error, up to +/-99 sec. Arrival time
error was simply the difference between the target arrival time and the arrivaltime computed from current speedand distance remaining. In an actual missionplanned speed would vary on eachnavigation leg,and soarrival time would beneeded for eachleg. In the simulation, planned speedwas constant acrosslegs,so only segment arrival time was displayed.
Waypoint Symbology - The Waypoint symbology consisted simply of a pennant
displayed at the geographical location of each waypoint and the altitude of the
aircraft. The pennants were maintained as moving models by the image
generation system but were displayed by the Silicon Graphics computer that
drove the helmet display. Each pennant was shaped like an arrow that pointed
toward the next waypoint (see Figure 5). Because the pennants were maintained
as part of the visual data base, all were displayed continuously, and their size
decreased with distance from the pilot's eye position.
30 33 N 03 0_
J I ! I i _ i J I I l * I
A
V
Figure 3. Basic helmet display symbology.
CDI Symbology - The CDI symbology mimicked a conventional, panel
mounted, course deviation indicator (CDI) (see Figure 6). A tail beneath the
compass lubber line pivoted to point toward the planned course. A carat (^)
indicated the heading to the current waypoint (as in the basic PNVS). A circle (o)
indicated the heading to the next waypoint. Both waypoint symbols edgelimited.
Alert Task Operator Interface - The pilot reacted to the alert symbol on the
helmet mounted display by performing a response task overlaid on the map
display using a keypad. The keypad was located on a console adjacent to thecollective lever and contained a button to allow the pilot to acknowledge the
alert along with a 10-key numeric pad. When the pilot acknowledged the alert,
the helmet display symbology ceased flashing, but the alert symbol remained
present. Acknowledgement also caused the task display to appear
6
superimposed on the map display. The task display consisted of three lines of
text. The top line was the word "LEFT" or "RIGHT," indicating the direction in
which the pilot was to enter numbers using the 10-key pad. The second line
showed five digits, selected randomly with replacement, which the pilot was to
enter. The third line showed five blanks, corresponding to the five digits to be
entered. As the pilot entered each correct digit, it was displayed in the
appropriate blank. Incorrect entries were ignored. Once the pilot entered the
fifth correct digit, both the map and helmet mounted displays returned to thenominal state.
30 33 N 03 04
A
V
N
Figure 4. Helmet display alert symbol
20 33 N 03 04
OKdN
<x/
> 23
Figure 5. Waypoint navigation symbol.
7
This task captured the salient aspects of a procedural task. These include a
structured sequence of actions, determination of required response, and making
the required response. For this task the sequence was first to acknowledge alert,
second to determine direction of digit input (left or right), and last to input five
digits. Unlike an actual flight procedure, this task had no consequence for the
flight.
2"q W 30 33
, I , ,0 i , _ i , , I , ,//
I
/X
i K < > 26 m,.V
Figure 6. "CDI" navigation symbology. Current waypoint symbolis edge limited on right.
Experimental Tasks - A standardized mission was developed, that consisted of
11 legs. The mission included four tasks: cross-country navigation, track
following, bob-up and reconnaissance for tanks. This mission is described in De
Maio and Hart (1999).
Procedure
Pilots participated in pairs. Each pilot performed one to three missions and thentook a break while the other pilot flew. The duration of each pair's simulation
period was four days. Missions lasted about 30 min. Following each mission the
pilot gave his workload and performance self-evaluation ratings without
receiving any feedback on his performance. Pilots received written instructionsexplaining the objectives of the research and the tasks that they would perform
(see Appendix B). They then performed familiarization flights until they were
ready to begin practicing the experimental tasks.
Pilots received paper maps that duplicated the cockpit map in order tofamiliarize themselves with the mission before hand. They were also allowed to
make a list of the waypoints for each mission to take into the cockpit. As the
pilots passed each waypoint, they were to depress the microphone switch and
state the waypoint name.
8
Three practice missions were provided. These missions had no tanks present,
and the pilots did not perform reconnaissance. The pilots flew practice missions
in each of the navigation display and alert conditions until they and the
experimenter felt that they were ready to proceed to the experimental trials.
Pilots gave no workload or self-evaluation ratings of the practice runs.
The order of presentation of the navigation display conditions for data collection
was balanced by a Latin Square (n = 12). Presentation of alert conditions was
balanced to control for first order effects. Pilots flew three to six experimental
missions per day for a total of 12 experimental missions.
Data collected for each experimental mission included workload ratings and
performance self-evaluations, mission time of waypoint passage reports, mission
time of reconnaissance reports, reconnaissance reports, and automatically
recorded flight performance data. The pilots gave workload ratings and
performance self-evaluations orally over the intercom, and the experimenter
transcribed them into a log, at the end of each mission.
It became clear during data collection that the bob-up task would not provide
usable data. Therefore a single task condition was added. The task was
performed in the cockpit on flight freeze. This condition is more like a classical
reaction time task in that the subject only monitored the helmet display for the
alert symbol. When the alert symbol appeared, the subject performed the alert
task just as in the flying condition. This condition is a logical extension of the
process of increasing the frequency of alerts. Unfortunately it was not included
in the original design. Some of the pilots had completed their tenure in the study
and were no longer available to participate in the reaction time task. Therefore
one of the experimenters and a simulation engineer served as subjects.
RESULTS
Problems with the timing of the bob-up task made it impossible to gather usable
data from that task. Only the track following and cross-country tasks provided
usable data for evaluating the alerts. The reaction time presentation of the alert
task also provided usable data, although the subjects in that task were not the
pilots who had participated in the simulation.
The first question was whether the alert and information manipulations provide
enough benefit to render average alert response latency shorter. An effect on the
mean might be expected if the manipulations had a beneficial effect on both long
and short response latencies. At least in the case of the full-screen alerting, the
author did not expect this to be the case. Rather the author expected that the full-
screen alert would provide a benefit primarily when the pilot's attention was
held strongly by another task, that is, in instances when very long response
latency might be expected. Therefor the alert manipulation should have moreeffect on the variance than on the mean.
On the other hand, one might expect that the information manipulation would
affect both long and short responses. Since the partial information alert affectsthe transfer of attention between displays, and not from the flying task to the
alert task, it should affect long and short responses equally.
The data analysis was a 2 (Info) X 2 (Alert) X 2 (Subject) analysis of variance forthe reaction time data and a 2 (Info) X 2 (Alert) X 8 (Subject) analysis of variance
for the simulation data. Mean response latencies are shown in the tables below.
Three responses were examined, as follows: the first response, to acknowledge
the alert; the second response, to enter the first digit; and the last response, to
enter the fifth digit. Analysis of variance tables are shown in Appendix C.
Three responses were use to describe the pilots' responses to the alert. Their first
response was to depress a button acknowledging the alert and turning off the
helmet display alert symbology. Their second response was to begin the digit
entry task. Their last response was to enter the final digit of the sequence.Intermediate digit entries were not examined. Table 2 shows the effect of the
alert and information manipulations on response latency in the reaction time
task. The manipulations produced reliable differences for the two digit entry
responses. The effect of the alert format was unexpected. The full-screen alert
produced a longer average latency for all three responses, and this differencewas statistically significant for the second and last responses. In the reaction
time task, the subjects' response was degraded by the full-screen presentation.The data were more in line with expectations for the partial information symbol.
Both of the responses involving digit entry were made more quickly when the
alert symbol provided partial information about the task. The initial response, to
acknowledge the alert, was non-significantly longer when partial information
was presented.
It would seem that any manipulation that increases the complexity of the alert,
either by adding information or by making the alert symbology more complex,
slows the initial response. One cannot say whether this result of this slowing
persists in subsequent responses. If so, the positive effect of partial information
on subsequent responses offsets this negative effect. On the whole, however, the
10
full-screen alert doesnot speed responding in a reaction time task, while the
partial information alert does speed responses that require use of thatinformation.
Table 2. Mean response latency (sec) for reaction time task. "*"
Indicates a reliably (p<0.05) shorter latency for that condition.
First
Response
Second
Response
Last
Response
Full-screen
Alert
2.23
4.10
7.98
Localized
Alert
2.20
3.89*
7.43*
Partial
Information
Symbol
2.27
3.58*
7.26*
No
Information
Symbol
2.18
4.38
8.12
Tables 3 and 4 show response latency data for the track following and cross-
country tasks, respectively. While none of the differences was statistically
reliable, there were some intriguing trends. First, the general trend for the partialinformation alert was similar to that obtained in the reaction time task, as was
expected, since the partial information affects task performance and not the
alerting quality of the symbology. Second, the full-screen alert showed a non-
significant trend toward faster responding. This trend was most pronounced for
the cross-country task, in which the alerts were least frequent. This trend shows
that the full-screen alert may be more effective in pulling attention from a
primary flying task to a rare event.
Table 3. Mean response latency (sec) for track following task. No
differences between means were statistically reliable.
First
Response
Second
Response
Last
Response
Full-screen
Alert
3.21
7.15
11.82
Localized
Alert
3.51
7.58
12.47
Partial
Information
Symbol
3.31
6.85
11.54
No
Information
Symbol
3.41
7.77
12.65
11
Theseresults suggest that both the full-screen alert and the partial information
alert might provide better alerting than do current, localized, uninformativealerts.
These results are not, however, as compelling as they might be. The magnitude
of the effects is at least as great as is seen in typical, single-task experiments on
partial information alerting (e.g., Rosenbaum, 1980; Reeve and Proctor, 1984). In
some cases the speed improvements exceeded one second, but the statistical
power is low. Part of the problem is that it was not possible to collect the large
amount of data usually collected in laboratory reaction time experiments.
Another part of the problem is that response latency data are not normally
distributed and that at least one of the manipulations may affect long latency
responses differentially. There is a need to perform analyses that focus more on
the long latency responses.
Table 4. Mean response latency (sec) for cross-country flight task.
No differences between means were statistically reliable. "t"
indicates p < 0.069.
First
Response
Second
Response
Last
Response
Full-screen
Alert
4.06-t
11.10
16.50
Localized
Alert
7.08
11.94
16.75
Partial
Information
Symbol
4.98
10.14
14.73
No
Information
Symbol
4.61
11.37
16.95
Differential effects on long latency responses affect the variance and skewness of
the distribution of responses much more than they do the mean. Appendix D
shows the distribution of responses for the various flight tasks and alert
conditions. All the distributions are skewed and that the degree of skewness
increases when alerts are less frequent..
The author used two statistical tools to examine the effect of the experimental
manipulation on the shape of the data distributions. The first, momental
skewness is a descriptive measure of the skewness of a distribution (Beyer, 1966,
p 3). It does not allow inferences about the magnitude of the difference in
skewness between two distributions, but it does allow us to quantify such
differences. The second, an F test for equality of variance is generally used to
ensure the suitability of data for analysis of variance (Hayes, 1963, p 351). In this
use, it measures the likelihood of a large deviation from normality, that is,
significant skewness. It has also been used to discriminate between differently
shaped data distributions (Hart and McPherson, 1976; Meyers, 1971).
12
Momental skewnesswas greater for the full-screen alert condition than for thelocalized alert condition for the first responseon the RT task but smaller for thelater responses(seeTable 5). Skewnesswas greater for the localized alertcondition in both the flying task conditions, reflecting the effect of fewer longlatency responsesin the full-screen alert condition. This effectmoderated forlater responses,and even reversed for the last responseon the cross-countryflight task.
Table 5. Momental skewnessasa function of alert type. Greater momentalskewnessindicates more long latency responses. The condition having lower
momental skewness,and fewer long latency responses,is shown in bold.
I st Response 2 nd Response Last Response
Full- Localized Full- Localized Full- Localized
screen screen screen
RT 5.94 1.56 1.39 1.83 3.25 5.06
TRK 4.29 5.06 2.41 3.07 1.17 2.44
CC 3.21 6.68 3.01 5.41 3.66 2.42
The partial information alert condition showed lower skewness across the board,
the only exception being the last response in the cross-country flight taskcondition (see Table 6). For the most part, the partial information alert did
reduce long latency responses.
Table 6. Momental skewness as a function of alert information content. Greater
momental skewness indicates more long latency responses. The condition
having lower momental skewness, and fewer long latency responses, is shown inbold.
I st Response 2 nd Response Last Response
Part Info No Info Part Info No Info Part Info No Info
RT 3.53 7.89 2.00
TRK 4.03 5.46 1.79
4.49 2.24 3.92
3.01 1.59 2.64
CC 3.29 7.48 3.16 4.82 6.20 1.72
The F-test for equality of variance showed that for the most part both the full-
screen alert (see Table 7) and the partial information alert (see Table 8) improved
latency performance, in the sense that they significantly reduced variance. The
effect of the partial information alert was weakest for the first response. The first
response showed more variability in two of the three task conditions. Variability
of the later responses was consistently reduced by the partial information alert.
13
Table 7. F- test for equality of variance of alert type data distributions.
"*" indicates p < 0.05. "_'1 st response data contained one very long latency;without it F (94, 102) = 4.05.
1 st Response 2 na Response Last Response
FvaR(nl, n2) RT 0.1 1.5" 2.9*
FvaR(nl, n2) Trk 2.4* 1.7" 1.7"
FvaR(nl, n2) CC 26.5* t 1.8" 1.2
Table 8. F- test for equality of variance of alert information data distributions.
"*" indicates p < 0.05; negative sign indicates larger partial information variance.
I st Response 2 nd Response Last Response
FVAR(nl, n2) RT -2.2* 1.6" 1.2
FvaR(nl, n2) Trk 1.7" 2.3*
-2.1" 1.7"FVAR(nl, n:) CC
Figure 7 presents a graphical summary of the latency data. There are trends
toward faster responding (mean latency), reduced skewness (difference between
mean and median), and reduced variability with the full-screen alert and with
partial alert information. The full-screen alert had its greatest impact on the first
response in the cross-country task. The effect diminished for later responses, and
the full-screen alert provided less benefit as the frequency of alerting increased.
The partial information content alert had its greatest effect on the later responses,
and may even have had a deleterious effect on the first response. Its effect was
roughly the same in all task conditions.
14
e• 3b
, DMean 1 31_'_ '
6 [] Median _5 6• Std Dev 5
2 nc
L,';::7,3 _ L_ _ ,o".o _.=.o1"
o 0 o
........ _1 m_l T ................. Trk Trk _ L_ zad
A. Alert Type - First Response
[] Mean
[] Median
• Stcl Dev
NOI_oRT Pitt Inlo NO Info Irk part _nfo NolnfoCC 13111 Info
F_T Trk (_
Info L_ll I Farm Talk
35
30
_S
_0 StO _,v
15 (le¢l
lo
5
o
D. Alert Information - First Response
15 40
[] Mean
[] Median _ 3o
La,an©y, • Std Dev(sac) S_ Dev
0 0
Scrla_ RT ScrHn Tr_ Screw1CC
Trk
Aloft Condition / Talk
B. Alert Type - Second Response
[] Mean12 ¸
• Std Dev9
I_c) (Nc)
s
o c o1 N • N • n• o o o • _o o
m rr_ oc
Inf= LIMI I Rl|hl T•U
E. Alert Information - Second Response
le 30 _e 30
DMean DMean
[] Median [] Median
,.,.. ,2 .StdDev __Ni] _==,. cc ....ste o "Stcl DevNo_T _ i _ 2°o
i s lacy) L'te_v s_ D_("¢) sic (sac) s (_c)
F_Sc_e_ tac=_T FU. SCmI_ L_ T_ _ull _ rkRT Dk _ NolnlORT • o I o P n fflt N IrdoCC pI_I inlo
m lr_ cc_art co_lmon I lllk
Into LIv•l I Flight Tllk
C. Alert Type - Last Response F. Alert Information - Last Response
Figure 7. Effects of experimental manipulations on mean response latency,
median response latency, and standard deviation of response latency.
15
DISCUSSION
The general trend in the data points to a small benefit from both the full-screen
alert and the partial information alert. The pattern of benefit varies for the two
manipulations. The full-screen alert provides its greatest benefit on the first
response, to acknowledge the alert, when alerts are infrequent. The partial
information alert provides its benefit on the execution of the alert task procedure,
independent of alert frequency. The measurement of these effects was somewhat
problematic, both because of the small size of the effects and because the effects
themselves significantly change the variance of the response latency distribution.
Yet these benefits come at virtually no cost, since they involve only modest
formatting and display of information readily available.
Full-Screen Alert
The full-screen alert has its largest effect on the time required for the pilot to
make an initial response to an infrequent alert. In this case the alert must pull
the pilot's attention from the flying task. Attracting the pilot's attention becomes
more difficult the less frequent the alerts. In the simulation alert frequency was
always very high compared to what might be expected in actual flight. In the
bob-up task, which did not produce usable data, three alerts were presented in
about 30 sec. In the track following task, alerts were presented about once per
minute. In the cross-country task one alert was presented in a 15 to 20 minute
segment. The benefit of the full-screen alert was greater in the cross-country
task. One might expect a substantially greater benefit in actual flight, when the
interval between alerts would typically be measured in hours or missions.
The magnitude of the full-screen alert effect declined over the course of
execution of the alert task. The initial benefit was swamped by other variability
in task performance. In actual flight the effect might be more persistent for a
number of reasons. The benefit might be larger; overlearned procedures should
be less variable; and performance of the alert procedure would be a higher
priority for the crew. So one might expect a substantial benefit from full-screen
alerting in actual flight.
A concern prior to the simulation was that the full-screen alert would be too
annoying and might even be disorienting. Intense strobe lights can degrademotor control and orientation. The author did not expect this to be a problem for
a variety of reasons. The luminance of the alert symbology was low. Usually
undesirable effects occur with more intense stimuli. The symbology was never
fully off, and the display brightness was relatively constant, so that the eye was
never in darkness. Finally the flash rate was well below that at which
disorientation occurs. In the event, the pilots reported no difficulty of any sortwith the full-screen alert.
Partial Information Alert
The purpose of the partial information alert was to forewarn the pilots about
what response they were to make. This forewarning should have facilitated task
execution (Rosenbaum, 1980; Leuthold et al, 1996). In fact the partial alert did
facilitate performance of the alert task procedure both in the reaction time
context and in the flying context. This benefit persisted throughout the full
16
execution of the task. Partial information did slow the initial response toacknowledge the alert somewhat, but this slowing was more than offset by theimprovement in task execution speed. The procedural task used was both lesscomplex and lessrehearsed than actual aircraft emergencyprocedures, and thatits impact on aircraft operation was nil.
17
REFERENCES
Aeronautical Design Standard 33D, Handling Qualities Requirements for
Military. Rotorcraft, U. S. Army Aviation and Troop Command, St. Louis, MO,1994.
Beyer, W. H. (ed.) Chemical Rubber Co. Handbook of Tables for Probability. and
Statistics, The Chemical Rubber Co., Cleveland, 1966.
De Maio, J. Attention Switching and the Time Required to Perform Two Tasks,
Unpublished Doctoral Dissertation, Arizona State University, 1976.
De Maio, J. and Hart, S. G. Situation Awareness and Workload Measures for
SAFOR, Joint Army - NASA Technical Memorandum, AFDD/TR-99-A-004;NASA/TM-1999-208795, 1999.
Hart, S. G. and McPherson, D. "Airline Pilot Time Estimation During Concurrent
Activity including Simulated Flight," Proceedings of the 47 _ Annual Meeting of
the Aerospace Medical Association, Bal Harbor, FL, 1976.
Hays, W. L. Statistics, Holt, Rinehart, and Winston, New York, 1963.
Hawkins, F. H. Human Factors in Flight, Ashgate, Brookfield, VT, 1997.
Hennessey, R. T. and Sharkey, T. J. Display of Aircraft State Information for
Ambient Vision Processing Using Helmet Mounted Displays, SBIR Phase I Final
Report, MTI-9701-962903-01, MTI, Los Gatos, CA, 1997.
Leuthold, H., Sommer, W., and Ulrich, R. "Partial Advance Information and
Response Preparation: Inferences from the Lateralized Readiness Potential,"
|ournal of Experimental Psychology: General, 125(3), 307 - 323, 1996.
Meyers, J. L. Fundamentals of Experimental Design, Allyn and Bacon, Boston,1971.
Reeve, T. G. and Proctor, R. W. "On the Advance Preparation of Discrete Finger
Responses," Tournal of Experimental Psychology: Human Perception and
Performance, 10(4), 541 - 553, 1984.
Rosenbaum, D. A. "Human Movement Initiation: Specification of Arm,
Direction, and Extent," [ournal of Experimental Psychology: General, 109(4), 444
- 474, 1980.
Woods, D. D., O'Brien, J. F., and Hanes, L. F. "Human Factors Challenges in
Process Control," in Handbook of Human Factors, (G. Salvendy, ed.), John Wiley
& Sons, New York, 1987.
18
APPENDIX A
MOTION GAINS
19
*** SLOW GAINS
GXS=.4
GYS=.8
GZS=.5
GPS=.05
GQS=.05GRS=.5
*** SLOW CORNER FREQUENCIES
OMEGXS=I.5OMEGYS=.6
OMEGZS=.3
OMEGPS=.5
OMEGQS=.7OMEGRS=.5
*** FAST GAINS
GXF=.4
GYF=.6
GZF=.9
GPF=.5
GQF=.8GRF--.4
*** FAST CORNER FREQUENCIES
OMEGXF=I.5
OMEGYF=.6
OMEGZF=I.4
OMEGPF=.85
OMEGQF=.85OMEGRF=.7
*** DAMPING RATIO
ZETAX=.707
ZETAY=.707
ZETAZ=.707
ZETAP=.707
ZETAQ=.707ZETAR=.707
*** SLOW AND FAST "SPEEDS" FOR INTERPOLATION OF SLOW AND FAST GAINS
VGFAST=10.
VGSLOW=0.
20
*** SLOW AND FAST "SPEEDS" FOR INTERPOLATION OF SLOW AND FAST CORNER
FREQUENCIES
VWFAST=10.
VGWLOW=0.
*** TC FORWARD PATH GAINS
GXTC=I.
GYTC=I.
*** TC FEEDBACK GAINS
GKTCFB=.I
*** RESIDUAL TILT CORNER FREQUENCIES. DAMPING RATIO, AND GAINS
*** CORNER FREQ. OF LARGE AXIS RT LOWPASS FILTER
OMEGLRT=2.
*** CORNER FREQ. OF SMALL AXIS RT LOWPASS FILTER
OMEGSRT=2.
*** DAMPING RATIO OF LARGE AXIS RT LOWPASS FILTER
ZETAR1=.707
ZETAR2=.707
*** ROLL AND PITCH RT GAINS
GPRT=I.
GQRT=I.
GKTCFB=0.5
21
Appendix B
Pilot Instructions
22
Introduction and Instructions to Experimental Pilots
This experiment is the start of a program of research, conducted jointly by the
Army and NASA, to develop principles for the presentation of symbolic
information on a helmet mounted display. The current work uses a production
AH-64 IHADSS, but the goal is to extend the work with an advanced color, wide
field of view, binocular display system. The research is conducted on NASA's
Vertical Motion Simulator, the largest motion based simulation in the world.
The research will examine presentation of two types of information on the
helmet display, alert information and navigation information. Alert information
will be presented in two modes. In a one mode a flashing letter presented at the
bottom of the HMD will indicated that you should come into the cockpit to
perform a procedural task. In the second mode the flashing letter will be
accompanied by flashing of all HMD symbology. Two navigation display
conditions will be used. In one waypoint data will be presented by markers onthe HMD that have been located and scaled to fit into the visual scene. In the
second an "instrument" located at the top of the HMD will present waypoint
information. The test conditions for both experimental questions will be
embedded in a simulated cross-country flight mission of about 20 minutesduration.
Cross-Country Flight
Prior to each flight, you will receive a map showing the route you are to fly. You
will maintain an altitude of 100 ft agl and an airspeed of 80 kt. At each waypoint
you will make a radio call to inform us that you have reached the waypoint. You
should make a list of the waypoints and headings to aid you in the cockpit.
Along the route you will perform two experimental flight tasks, that are not part
of the cross-country flight. These tasks are a track following tasks and a bob-up.
Your flight time from the take-off to the track and from the track to the bob-up
position will be specified for each mission. You will also be asked to reconnoiter
two areas of interest along each route and report back the number of tank
platoons in each area. Performance specifications are given in the table at theend of this document.
Track Following Task
You will follow a track on the ground maintaining a comfortable and an altitudeof 100 ft AGL.
Bob-up Task
You will establish a stable hover at 10 ft AGL in position in front of the bob-up
tree. At this point make your radio call to tell us that you have reached the
waypoint. This call will initiate the task. You will then climb to and altitude of50 ft AGL and hover for 10 sec. Mark the start and end of the hover and the end
of the task by pressing the "xmit" switch. Performance specifications are givenin the table at the end of this document.
Procedural Tasks and Alert Displays
23
We have developed a laboratory task which is intended to demand your
attention as would an inflight procedure. This task will be presented on the left
panel CRT, which normally displays the moving map. The task requires you to
enter five digits on the numeric keypad located on the side console. The word
"Left" or "Right" displayed above the number indicates whether they are to be
entered from left to right or from right to left. The system will only respond to
correct entries, which will be displayed on the panel. After you have entered all
five digits, the map display will reappear.
An HMD alert will signal when you are to perform this task. This signal will
consist of a flashing letter at the bottom of the HMD. In on condition, only this
letter will flash. In a second condition, all HMD symbology will flash. The
symbol may provide some information about the task to be performed. In one
condition, the letter, "L" or "R" will indicate left-to-right or right-to-left entry. In
a second condition, the letter "N" will indicate only that the number entry task is
to be performed. Prior to beginning the number entry task, you must "clear" the
alert by pushing the "Cancel" button next to the numeric pad.
The sequence of events for this task is as follows:
1. HMD flashes and number entry task replaces moving map
2. Pilot "clears" alert by pushing "Clear" button
3. Pilot enters five numbers on keypad
4. Panel display reverts to moving map.
Navigation Displays
There are two HMD navigation displays.
A "lollipop" display indicates the location of the current waypoint by a symbol
floating above its location. The symbol will point left or right depending on the
direction of turn required after passing the waypoint. A timer will track your on-
time performance. A time of arrival error counter located in the upper right
portion of the HMD tells how many seconds early (+) or late (-) you are.
A "CDI" display shows your deviation from the route, heading to the current
waypoint, heading to the next waypoint, and on-time status. Deviation is
indicated by a pointer located below the lubber line. This pointer indicates the
direction to the course. Maximum scaled deviation is 2000 ft when the pointer is
60 degrees from vertical. The current waypoint is indicated by a carat on the
compass scale. The next waypoint is indicated by a circle on the compass scale.
These symbols edge limit if the waypoint is off scale. On-time status is indicated
as on the "Lollipop" display.
Performance Criteria and Workload
You will be asked to evaluate your performance against the criteria in the table
below. You will be asked to rate your workload on six phases of the mission.
Your ratings will be made on a scale from I to 100. Ratings will be made for
"Input," that is, gathering information, "Central," that is, thinking about the task,
"Output," that is, making control inputs or other actions, and "Time" pressure.
24
Recon
Navigation
Accuracy
(% tanks
detected)
Timeliness
(report timeafter first
detection
Accuracy(maximumdeviation
from
course)
D
(Desired)
>90%
<20 sec
<100
A
(Acceptable)
75% - 90%
20 - 40 sec
100-200
O
(Outside
Acceptable)
<75%
>40 sec
>200 ff
Timeliness +/- 10 sec of +/- 20 sec of >+/- 20 sec of
(at track assigned time assigned time assigned timeand bob-up)
Height +/- 3 ft +/- 6 ft >+/- 6 ft
Bob-up Time +/- 4 sec +/- 6 sec >+/- 6 sec
Position +/- 6 ft +/- 10 ft >+/- 10 ft
Performance criteria.
25
APPENDIX C
DISTRIBUTION OF REACTION TIMES
26
Ranked 20Di,,t from
Md
40
RT 1st Resp
-- Full Screen
O Mean
O 10 20Momental
Skewness = 5.94 Latency (sac)
3O
Ranked 20
Dlst from
Md 40
RT Ist Rasp
_Localized
O Mean
0 10 2OMomentalSkewness = 1.56 Latency (sac)
3O
Track 1st Resp
oA20
Ranked 40Diet from _ Full Screen
Md 60
100 -
0 10 20 30Momental
Skewness = 4.29 Latency (eoc)
Track 1st Resp
oL2O
Ranked 40Diet from
Md 60 _ Localized
80 O Mean
0 10 20Momental
Skewness = 5.06 Latency (sac)
3O
Cross Country 1st Resp
Ranked 10
Diet from
Md 20 -- Full Screen
30 Mean
40
0 10 20
Momental Latency (sac)Skewness = 3.21
3O
Cross Country 1st Resp
Ranked 10
Diet fromM d 20 Localized
0 10 20 30
Momental Latency (sac)Skewness = 668
Distribution of Response Latencies for First Response as a Function of Display.
Positive skewness is evident in the tail of long latency responses and in the
displacement of the mean relative to median. Mean will be displaced to a higher
valued when distribution is more highly skewed. Greater value of momental
skewness statistic indicates greater skewness.
27
20Ranked Diet
from Md40
0
Momental
Skewness = 183
RT 2nd Rasp RT 2nd Resp
20
Ranked D_st
-- Full Screen from Md 40
O Mean
10 20 30 0 10Momental
Latency (Bee) Skewness = 1.39 Latency (aac)
Localized
0 Mean
20 30
0
2ORanked
Dist 40
from 60
Md 80
100
0
Momental
Skewness = 3,07
TRK 2nd Resp TRK 2nd Resp
t_ Ranked 20
Disf 40 •
from 60
O Mean Md 80 '
100 '
10 20 30 0 10 20
Momental Latency (lee)Latency (lee) Skewness = 2.41
-- Localized
Mean
3O
Cross-country 2nd Resp Cross-country 2nd Resp
0 0/%Ranked 10 Ranked 10 --LocalizedOist Full-screen Dist
Md 30 _ Md 30
40
0 10 20 30 0 10 20 30
Momental Latency (leo)Momental Latency (aac) Skewness = 3,01Skewness = 5.41
Distribution of response latencies for second response as a function of display.
Positive skewness is evident in the tail of long latency responses and in the
displacement of the mean relative to median. Mean will be displaced to a highervalued when distribution is more highly skewed. Greater value of momental
skewness statistic indicates greater skewness.
28
Ranked 20
Dist
from 40Md
0Momental
Skewness = 366
RT 3rd Resp RT 3rd Resp
Ranked 20
-- Full-screen Diet _ LocalizedO Mean from 40
Md O Mean
10 20 30 40 50 0 10 20 30 40Momental
Letency( aec) Skewness =224 Latency (see)
5O
0
Ranked 20
Dist 40
fern 60
Md80
100
0Momental
Skewness = 1.17
TRK 3rd Resp TRK 3rd Resp
Ranked 20
Dist 40
_ Full-screen from 60 Localized
y "L,_._ _ Mean Md 80 / _ _an
10 20 30 40 50 0 10 20 30 40 50Momenta_
Latency (sec) Skewness = 2.44 Latency (lec)
Cross-country 3rd Resp Cross-country 3rd Resp
0 \ ° ARanked 10 Ranked 10Full-screen _ Localized
Olst 20 - Dist
Md 30 Md 30
40
0 10 20 30 40 50 0 10 20 30 40 50MomentalSkewness = 325 Latency (see) Momental
Skewness = 5.06 Latency (aec)
Distribution of response latencies for third response as a function of display.
Positive skewness is evident in the tail of long latency responses and in the
displacement of the mean relative to median. Mean will be displaced to a highervalued when distribution is more highly skewed. Greater value of momental
skewness statistic indicates greater skewness.
29
Ranked 20
Diet
from 40Md
0
Momental
Skewness = 329
RT 1st Reap
No Info
O Mean
10 20 30
Latency (lec)
RT 1st
Rankad 20
Dist
from 40Md
0
Momental
Skewness = 7.48
Reap
O
10 20
Latency (sac)
-- Part Info
Mean
3O
TRK 1st Reap
20 "
Ranked 40
Dist 60from _No Info
Md 80 O Mean
100
0 10 20
Momental Latency (sac)Skewness = 546
30
TRK 1st Reap
oL20Ranked
Dist 40
from 60Md
80
0 10 20
Momental Latency (aec)Skewness = 4.03
_Pa_lnfo
O M_n
3O
Cross-Country 1st Reap
0
Ranked 10
Dist 20from
Md 30
4O
0
Momental
Skewness = 7.89
/_ _ No InfoO Mean
10 20
Latency (lee)
Cross-Country 1st Reap
Ranked10
Distfrom 20 _Part Info
Md 30 J "_=..,-.___..__.__ Mean
30 0 10 20 30
Momental Latency (sac)Skewness = 3.53
Distribution of response latencies for first response as a function of information.Positive skewness is evident in the tail of long latency responses and in the
displacement of the mean relative to median. Mean will be displaced to a highervalued when distribution is more highly skewed. Greater value of momental
skewness statistic indicates greater skewness.
3O
RT 2nd Resp RT 2nd Resp
Ranked 20Disl
from 40Md
0Momental
Skewness = 4.82
oLRanKe0 20Dist
_No Info from 40 _Part InfoO Mean M0
O Mean
10 20 30 0 10 20 30Momental
Latency (lee) Skewness = 3.16 Latency (aec)
0
20Ranked
Disf 40
from 60
MO 80
100
0Momental
Skewness = 3.01
Trk 2nd Resp Trk 2nd Resp
20Ranke0
Dist 40
_No tnfo from 60 _Part InfoMO
o Mean 80 / "_..,._ O Mean
10 20" 30 0 10 20 30
Latency (sic) MomentalSkewness = 1.79 Latency (leo)
Cross-Country 2nd Resp
o
Flankeci 10
Dist20
fromMd
30
4O
0Momental
Skewness = 4.49
Cross-Country 2nd Resp
Ranke0 10
Dist_No Info from 20 Part Info
/ /_o _
10 20 30 0 10 20 30Momental
Latency (sac) Skewness=2.OO Latency (aac)
Distribution of response latencies for first response as a function of information.
Positive skewness is evident in the tail of long latency responses and in the
displacement of the mean relative to median. Mean will be displaced to a highervalued when distribution is more highly skewed. Greater value of momental
skewness statistic indicates greater skewness.
31
RT 3rd Resp RT 3rd Resp
°LRanked 20 Ranked 20
Dist Dist
from 40 ' O Mean from 40Md Md
0 10 20 30 40 50 0 10 20Momental MomentalSkewness = 366 Latency(sec) Skewness = 2.24 Latency (sec)
Localized
0 Mean
30 40 50
TRK 3rd Resp TRK 3rd Resp
Renked20 20
Dist 40 Dist 40
fern 60 _Full-screen from 60 _Localized
Md 80 0 Mean Md 80 / __..._an100
0 10 20 30 40 50 0 10 20 30 40 50Momental MomentalSkewness = 1.17 Latency (sac) Skewness = 2,44 Latency (sec)
Cross-country 3rd Resp Cross-country 3rd Resp
o o ARanked 10 _Full-screen Ranked 10 -- LocalizedOist 20 Dist
Md 30 Md 30
40
0 10 20 30 40 50 0 10 20 30 40 50Momental MomentalSkewness = 3.25 Latency (sac) Skewness = 5.06 Latency (sec)
Distribution of response latencies for first response as a function of information.
Positive skewness is evident in the tail of long latency responses and in the
displacement of the mean relative to median. Mean will be displaced to a higher
valued when distribution is more highly skewed. Greater value of momental
skewness statistic indicates greater skewness.
32
APPENDIX D
ANALYSIS OF VARIANCE TABLES
33
General Linear Models Procedure
Dependent Variable: Time to First Response, RT Task
Source DF Sum of Squares Mean Square
Model 3 2.228 0.743
Error 209 34.238 0.164
Corrected Total 212 36.466
R-Square C.V. Root MSE
0.061 18.13 0.405
Source DF Type III SS Mean Square
PILOT 1 1.792 1.792
ALERT 1 0.0026 0.0026
INFO 1 0.499 0.499
General Linear Models Procedure
Dependent Variable: Time to Second Response, RT Task
Source DF Sum of Squares Mean Square
Model 3 54.31 18.103
Error 210 169.598 0.807
Corrected Total 213 223.908
R-Square C.V. Root MSE
0.24 22.51 0.899
Source DF Type III SS Mean Square
PILOT 1 8.135 8.135
ALERT 1 3.604 3.604
INFO 1 40.962 40.962
F Value
4.53
C Mean
2.232
F Value
10.94
0.O2
3.05
F Value
22.42
C Mean
3.992
F Value
10.07
4046
50.72
Pr>F
0.004
Pr>F
0.0011
0.90
0.082
Pr> F
0.0001
Pr> F
0.0017
0.0358
0.0001
34
General Linear Models Procedure
Dependent Variable: Time to Last Response,RT Task
Source DF Sum of Squares Mean Square
Model 3 118.651 39.550
Error 210 858.143 4.086
Corrected Total 218 976.793
R-Square C.V. Root MSE
0.12 26.16 2.021
Source DF Type III SS Mean Square
PILOT 1 31.998 31.998
TASK 1 23.844 23.844
PILOT*TASK 1 57.392 57.392
General Linear Models Procedure
Dependent Variable: Time to First Response, Cross-Country
Source DF Sum of Squares Mean Square
Model 9 349.49 36.832
Error 157 2439.88 15.541
Corrected Total 166 2789.36
R-Square C.V. Root MSE
0.12 84.23 3.94
Source DF Type III SS Mean Square
PILOT 7 285.71 31.998
TASK 1 52.18 23.844
PILOT*TASK 1 0.25 57.392
F Value
9.68
C Mean
7.726
F Value
7.83
5.84
14.04
F Value
2.50
C Mean
4.68
F Value
2.63
3.36
0.02
Pr> F
0.0001
Pr> F
0.0056
0.0166
0.0002
Pr> F
0.0107
Pr> F
0.0136
0.0688
0.8996
35
Form Approved
REPORT DOCUMENTATION PAGE OM8No 0z04-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources,gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of thiscollection of information, including suggestions for reducing this burden, to Washington HeadQuarters Services, Directorate for information Operations and Reports, 1215 JeffersonDavis Highway, Suite 1204, Arlington, VA 22202-4302, end to the Office of Management and Budget, Paperwork Reduction Pro act (0704-0188}, Washington, DC 20503
1. AGENCY USE ONLY (Leave blenk) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
September 2000 Technical Memorandum4. TITLE AND SUBTITLE 5. FUNDING NUMBERS
Evaluation of Helmet Mounted Display Alerting Symbology
6. AUTHOR(S)
Joe De Maid
7. PERFORMINGORGANIZATIONNAME(S)ANDADDRESS(ES)
Aeroflightdynamics Directorate
U.S. Army Aviation and Missile CommandAmes Research Center
Moffett Field, CA 94035-1000
9. SPONSORING/MONITORINGAGENCYNAME(S)ANDADDRESS(ES)
National Aeronautics and Space Administration
Washington, DC 20546-0001U.S. Army Aviation and Missile Command
Redstone Arsenal, AL 35898-5000
581-31-22
8. PERFORMING ORGANIZATION
REPORT NUMBER
A-00V0023
10. SPONSORING/MONITORING
AGENCY REPORT NUMBER
NASA/TM-2000-209603
AFDDfFR-00-A-009
11. SUPPLEMENTARY NOTES
Point of Contact: Joe De Maid, Ames Research Center, MS 210-5, Moffett Field, CA 94035-1000
(650) 604-6074
12a. DISTRIBUTION/AVAILABILITY STATEMENT
Unclassified -- Unlimited
Subject Category 53 Distribution: Standard
Availability: NASA CASI (301) 621-0390
12b. DISTRIBUTION CODE
13. ABSTRACT (Maximum 200 words)
Proposed helicopter helmet mounted displays will be used to alert the pilot to a variety of conditions,
from threats to equipment problems. The present research was performed under the NASA SAFOR, sup-
ported by a joint Army/NASA research agreement. The purpose of the research was to examine ways to
optimize the alerting effectiveness of helmet display symbology. The research used two approaches to
increasing the effectiveness of alerts. One was to increase the ability of the alert to attract attention by using
the entire display surface. The other was to include information about the required response in the alert
itself. The investigation was conducted using the NASA Ames Research Center's six-degree-of-freedom
vertical motion simulator (VMS) with a rotorcraft cockpit. Helmet display symbology was based on the
AH-64's pilot night vision system (PNVS), cruise mode symbology. A standardized mission was developed,
that consisted of 11 legs. The mission included four tasks, which allowed variation in the frequency of
alerts. The general trend in the data points to a small benefit from both the full-screen alert and the partial
information alert.
14. SUBJECT TERMS
Helmet display, Pilot alerting, Partial information, MHD
17. SECURITY CLASSIFICATION
OF REPORT
Unclassified
NSN 7540-01-280-5500
18. SECURITY CLASSIFICATION
OF THIS PAGE
Unclassified
19. SECURITY CLASSIFICATION
OF ABSTRACT
Unclassified
lS. NUMBER OF PAGES
4116. PRICE CODE
A03
20. LIMITATION OF ABSTRACT
Standard Form 298 (Rev. 2-89)Prescribed by ANSI Std Z3e-18
;->98-102
NASA/TM-2000-209603
AFDD/TR-00-A-009
Evaluation of Helmet Mounted Display Alerting Symbology
Joe De Maio
September 2000
The NASA STI Program Office... in Profile
Since its founding, NASA has been dedicated to the
advancement of aeronautics and space science. The
NASA Scientific and Technical Information (STI)
Program Office plays a key part in helping NASA
maintain this important role.
The NASA STI Program Office is operated by
Langley Research Center, the Lead Center forNASA's scientific and technical information. The
NASA STI Program Office provides access to the
NASA STI Database, the largest collection of
aeronautical and space science STI in the world.
The Program Office is also NASA's institutional
mechanism for disseminating the results of its
research and development activities. These results
are published by NASA in the NASA STI ReportSeries, which includes the following report types:
TECHNICAL PUBLICATION. Reports of
completed research or a major significant phase
of research that present the results of NASA
programs and include extensive data or theoreti-
cal analysis. Includes compilations of significantscientific and technical data and information
deemed to be of continuing reference value.
NASA's counterpart of peer-reviewed formal
professional papers but has less stringent
limitations on manuscript length and extent
of graphic presentations.
• TECHNICAL MEMORANDUM. Scientific and
technical findings that are preliminary or of
specialized interest, e.g., quick release reports,
working papers, and bibliographies that containminimal annotation. Does not contain extensive
analysis.
CONTRACTOR REPORT. Scientific and
technical findings by NASA-sponsored
contractors and grantees.
CONFERENCE PUBLICATION. Collected
papers from scientific and technical confer-
ences, symposia, seminars, or other meetings
sponsored or cosponsored by NASA.
SPECIAL PUBLICATION. Scientific, technical,
or historical information from NASA programs,
projects, and missions, often concerned with
subjects having substantial public interest.
TECHNICAL TRANSLATION. English-
language translations of foreign scientific and
technical material pertinent to NASA's mission.
Specialized services that complement the STI
Program Office's diverse offerings include creating
custom thesauri, building customized databases,
organizing and publishing research results.., even
providing videos.
For more information about the NASA STI
Program Office, see the following:
• Access the NASA STI Program Home Page at
http://www.sti.nasa.gov
• E-mail your question via the Internet to
help@ sti.nasa.gov
• Fax your question to the NASA Access Help
Desk at (301) 621-0134
• Telephone the NASA Access Help Desk at
(301) 621-0390
Write to:
NASA Access Help DeskNASA Center for AeroSpace Information
7121 Standard Drive
Hanover, MD 21076-1320
Recommended